I have a lot of regions on a single track, which i would like to apply the processing from the respective channel to. I only find a way to either combine them with processing or bounce them with processing. But the bounced regions will appear in the sources list and i would need to reimport them again. Combining would keep the sounds in the right place but it would make one region out of all these individual ones.
Is there a way to achieve what i want? (Apply processing to multiple regions at once and keep them in the same position/length)
What is your actual goal? Are you attempting to reduce the processor load by rendering processing to a file so that you can disable the processors on that track?
No, i am creating a sample library. I did cut a recording into the samples i want to continue working with. Now i want to highpass filter them and after that normalize them. So that i can later adjust compressor settings more easy. I noticed that it would have been smarter to first apply the highpass and then cut it but now it’s too late. 
I still don’t understand why you want to bounce at that stage. Since you mention re-importing the regions that seems to imply you are not finished working with them. Why not just perform the high-pass filtering in real time as you are continuing to work with the regions?
Why do you say that? I assume you started with one large region, and are splitting into multiple small regions. Why would it matter if you applied a filter before or after splitting the regions? You do mention channel processing, by which I assume you mean plugins on a track. Since the plugins are in the track signal chain the number of regions the audio recording is split into has no affect.
Normalization is applied to each single region. When i filter out the low frequencies after normalization the regions are not normalized anymore. I want the filtered regions to all have the same peak levels.